HK1206574B - X-ray reduction system - Google Patents

Description

X-ray reduction system

Technical Field

The present invention relates to the field of fluoroscopy, and more specifically to the field of controlling the amount of X-ray radiation during fluoroscopy.

Cross reference to related patent applications

This patent application claims priority to and is related to U.S. provisional patent application No. 61/606,375, filed on 3/2012, which is incorporated herein by reference in its entirety.

Background

In a typical fluoroscopy system, an X-ray tube generates X-ray radiation over a relatively wide solid angle. To avoid unnecessary radiation to the patient and medical team, a collimator of X-ray absorbing material, such as lead, is used to block the excess radiation. In this way only the necessary solid angle of useful radiation leaves the X-ray tube to illuminate the necessary elements.

Such a collimator is usually used in a static mode, but it may have different designs and X-ray radiation geometries. The collimator may be set manually or automatically using inputs including, for example, the size of the organ in the environment involved.

In fluoroscopy, the situation is more dynamic than with single radiation X-rays. X-ray radiation is emitted over a relatively long period of time, and medical personnel are often required to stand beside the patient, and thus close to the X-ray radiation. Therefore, it is desirable to provide a method to minimize exposure to the medical team. Methods of reducing the intensity of X-ray radiation have been proposed in which the reduction in the signal-to-noise ratio (S/N) of the X-ray image is compensated by digital image enhancement. Other approaches propose that the collimator limits the solid angle of X-ray radiation to a fraction of the image intensifier area and move the collimator to swap the entire input area of the image intensifier, with a region of interest (ROI) illuminated more than the rest of the area. Thus, the ROI receives a sufficiently high X-ray radiation to produce a high S/N image, while the remainder of the image is irradiated with a low X-ray intensity, providing a relatively low S/N image. The size and location of the ROI can be determined in a variety of ways. For example, it may be a fixed region in the center of the image, or it may automatically center around the most active region of the image, this activity being determined by real-time image analysis of a series of photographic images received from a video camera of a fluoroscopy system.

Disclosure of Invention

According to a first aspect of the present invention there is provided an X-ray system comprising an X-ray source, a single substantially circular collimator, a detector, a display and means for rotating the collimator about an axis substantially perpendicular to the plane of the collimator; wherein the collimator is constructed of a region that is substantially opaque to X-ray radiation and a region that is transparent to X-rays.

The detector may comprise means for integrating the signal during a rotation of said collimator around 360 degrees (exposure period), said system further comprising means for reading a frame comprising pixel values from said detector.

The means for reading may comprise means for reading the values comprising the pixels at the end of each exposure period.

The means for reading may comprise means for reading the values comprising the pixels at the end of an integer number of exposure periods.

The means for reading may comprise means for reading the values comprising the pixels before the end of each exposure period.

The number of frames read during the exposure period may be an integer.

The system may further include means for calculating gain and offset corrections for each frame, and means for calculating a normalization factor for each frame based on collimator shape, velocity, and position.

The system may also be configured to generate an exposure image by correcting and accumulating all frames during an exposure period.

The system may also be configured to generate an exposure image by correcting and accumulating the last read N frames after each frame is read.

The system may also be configured to generate an exposure image by correcting and accumulating all of the exposed pixels in the frame read during the exposure period.

The system may also be configured to generate an exposure image by correcting and accumulating all of the exposed pixels in the last read N frames after reading each frame.

The means for calculating the normalization factor for each frame may comprise means for multiplying the pixel values by a theoretical factor that compensates for DPP differences.

The means for calculating a normalization factor for each frame may comprise means for obtaining a calibration frame and means for calculating a calibration factor for each pixel from the calibration frame.

The calibration frame may comprise an average of a plurality of frames.

The calibration frame may include one frame captured when X-ray radiation is on and one frame captured when X-ray radiation is off.

The means for calculating the normalization factor for each frame may comprise means for calculating a bilinear calibration.

The system may further comprise means for generating an exposure image from said corrected frame and means for refreshing said exposure image.

Said means for refreshing may comprise means for using different refresh rates for different regions of said image.

The means for reading may comprise means for sequentially accessing the detector and reading the entire frame therefrom.

The transmissive region may be a combination of a circular hole concentric with the collimator center and a hole of a shape of a portion of a circle concentric with the collimator center of rotation and spanning an angle.

The collimator may further comprise a balancing factor.

The frame may include pixels associated with the circular aperture receiving a first dose of X-ray radiation and pixels associated with a collimator area around the circular aperture receiving a second dose of X-ray radiation, the second dose including a portion of the first dose proportional to a ratio of a fan angle to 360 degrees.

The means for reading may comprise means for randomly accessing the detector and reading pixels therefrom.

The means for reading may be configured to read pixel values from a first fully exposed sector adjacent to a currently exposed sector and reset the pixels after reading.

The angular span of said first sector may be chosen such that the time required to read and reset pixels within said first sector does not exceed the time required for said collimator to rotate the same angular distance.

The system may also be configured to reset pixel values in a second sector to be exposed, the second sector being adjacent to the currently exposed sector.

The collimator may further comprise synchronization means for synchronizing the detector with the collimator rotation.

The synchronization means may comprise a protrusion through the light sensor configured on the collimator.

The synchronization means may comprise an encoder.

The means for rotating the collimator may include a first pulley mounted on top of the collimator at a concentric position of the collimator; a second pulley mounted on the motor; a belt connecting the first pulley to the second pulley; a V-shaped circular track concentric with the collimator; and three wheels connected with the V-shaped grooves of the rails, the rotating shafts of the three wheels being fitted on an annular stationary part fixed to a reference frame of the X-ray tube.

The belt may be selected from the group consisting of: flat belt, circular belt, V-shaped belt, multi-grooved belt, ribbed belt, film belt and timing belt.

The means for rotating the collimator may comprise a gear transmission.

The gear transmission may be selected from the group consisting of: straight, helical, bevel, hypoid, crown and screw gears.

The means for rotating the collimator may comprise a high friction rotating surface post in direct contact with an edge of the collimator.

The collimator may comprise a fixed aperture.

The collimator may comprise a variable aperture.

The system may comprise means for assembling two concentric fixed aperture collimators and means for rotating one of the two collimators relative to the other.

The system may comprise means for rotating each of the two collimators independently.

The system may comprise means for rotating each of the two collimators at a different speed.

The system may comprise means for rotating the collimator at a variable speed.

The collimator may comprise an aperture shape designed to provide two regions of two different radiation Doses Per Pixel (DPP) at a fixed rotation speed.

The collimator may comprise a qualitative exposure arrangement for providing DPP of different levels at different distances from the collimator center.

The system may further comprise an eye tracker configured to track the gaze of an operator, thereby determining a region of interest (ROI) and controlling the collimator accordingly.

According to a second aspect of the present invention there is provided an X-ray system comprising an X-ray source, a single substantially circular collimator, a camera, a detector and a display, means for moving said collimator in a plane substantially parallel to the plane of said collimator; wherein the collimator comprises a central aperture allowing all radiation to pass through, an outer ring for reducing the amount of radiation passing through in dependence on the material and the thickness of the material, and an inner ring between the central aperture and the outer ring, the thickness of the inner ring varying as a function of the distance from the center, starting at zero on the side of the central aperture and ending at the thickness of the outer ring on the side of the outer ring.

The detector may be configured to integrate the signal of each frame captured by the camera, the system further configured to read frames comprising pixels from the detector.

The system may be configured to read a frame at the end of each frame captured by the camera.

The system may be further configured to calculate a gain per frame and an offset correction per frame, and calculate a normalization factor per frame from a different DPP for each of the aperture, the outer ring, and the inner ring of the collimator.

The system may be further configured to calculate the normalization factor for the inner annulus by dividing the inner annulus into a plurality of annuli and assigning one theoretical DPP to each of the plurality of annuli as a function of distance from the central aperture.

The system may further comprise an eye tracker configured to track the gaze of an operator, thereby determining a region of interest (ROI) and controlling the collimator accordingly.

According to a third aspect of the present invention there is provided a method of enhancing a displayed exposure image in an X-ray system, the system comprising an X-ray source, a single substantially circular collimator, a detector, a display and means for rotating said collimator about an axis substantially perpendicular to the plane of said collimator; the collimator is constructed of a region substantially opaque to X-ray radiation and a region transmissive to X-rays, comprising: said detector capturing images from said X-ray source; the detector integrates the signal during the rotation of the collimator around 360 degrees (exposure period); reading a frame comprising pixel values from said detector; calculating a gain and offset correction for each frame; calculating a normalization factor for each frame based on the collimator shape, velocity, and position; generating an exposure image from said corrected frame; and refreshing the exposed image.

The reading may comprise reading the pixel values at the end of each exposure period.

The reading may comprise reading the pixel values at the end of an integer number of exposure periods.

The reading may comprise reading the pixel values before the end of each exposure period.

The number of frames read during the exposure period may be an integer.

The generating of the exposure image may include correcting and accumulating all frames during the exposure period.

The generating of the exposure image may include correcting and accumulating the last read N frames after reading each frame.

The generating of the exposure image may include correcting and accumulating all of the exposure pixels in the frame read during the exposure period.

The generating of the exposure image may include correcting and accumulating all of the exposure pixels in the last read N frames after reading each frame.

The calculating of the normalization factor for each frame may include multiplying the pixel values by a theoretical factor that compensates for the DPP difference.

The calculating a normalization factor for each frame may include obtaining a calibration frame and calculating a calibration factor for each pixel based on the calibration frame.

The calibration frame may comprise an average of a plurality of frames.

The calibration frame may include one frame captured when X-ray radiation is on and one frame captured when X-ray radiation is off.

The calculating the normalization factor for each frame may include calculating a bilinear calibration.

The refreshing may comprise using different refresh rates for different regions of the image.

The reading may comprise sequentially accessing the detector and reading the entire frame therefrom.

The transmissive region may be a combination of a circular hole concentric with the collimator center and a hole of a shape of a portion of a circle concentric with the collimator center of rotation and spanning an angle; wherein said frame may comprise pixels of said circular aperture receiving a first dose of X-ray radiation and pixels of a collimator area surrounding said circular aperture receiving a second dose of X-ray radiation, said second dose comprising a portion of said first dose, said portion being proportional to a ratio of a fan angle and 360 degrees; the reading may include randomly accessing the detector and reading pixels therefrom.

The reading may include reading pixel values from a first fully exposed sector adjacent to a currently exposed sector and resetting the pixels after reading.

The angular span of the first sector may be selected such that the time required to read and reset pixels within the first sector does not exceed the time required for the collimator to rotate the same angular distance.

The method may further comprise resetting pixel values in a second sector to be exposed, the second sector being adjacent to the currently exposed sector.

The method may further comprise tracking the gaze of the operator, thereby determining a region of interest (ROI) and controlling the collimator accordingly.

According to a fourth aspect of the present invention there is provided a method of enhancing a displayed exposure image in an X-ray system, the system comprising an X-ray source, a single substantially circular collimator, a camera, a detector and a display, means for moving said collimator in a plane substantially parallel to the plane of said collimator; wherein the collimator comprises a central aperture allowing all radiation to pass through, an outer ring for reducing the amount of radiation passing through in dependence on the material and the thickness of the material, and an inner ring between the central aperture and the outer ring, the thickness of the inner ring varying as a function of the distance from the center, starting at zero on the side of the central aperture and ending at the thickness of the outer ring on the side of the outer ring, comprising: the detector integrates the signal of each frame captured by the camera; reading a frame comprising pixels from said detector; calculating a gain for each frame and an offset correction for each frame; calculating a normalization factor for each frame from a different DPP for each of the aperture, the outer ring and the inner ring of the collimator.

The reading may comprise reading a frame at the end of each frame captured by the camera.

The calculating the normalization factor for the inner annular ring may include dividing the inner annular ring into a plurality of rings and assigning one DPP theoretical value to each of the plurality of rings according to a distance from the central aperture.

The method may further comprise tracking the gaze of the operator, thereby determining a region of interest (ROI) and controlling the collimator accordingly.

Drawings

For a better understanding of the invention, and to show how the same may be carried into effect, reference will now be made, purely by way of example, to the accompanying drawings.

With specific reference now to the drawings in detail, it is stressed that the particulars shown are by way of example and for purposes of illustrative discussion of the preferred embodiments of the present invention only, and are presented in the cause of providing what is believed to be the most useful and readily understood description of the principles and conceptual aspects of the invention. In this regard, no attempt is made to show structural details in more detail than is necessary for a fundamental understanding of the invention. It will be apparent to those skilled in the art how the various forms of the invention may be practiced in practice with reference to the accompanying drawings and description. In the drawings:

FIG. 1A is a simplified schematic diagram of an example of an arrangement of a fluoroscopic clinical environment and system;

FIG. 1B is an illustration of an example of a layout of the system of FIG. 1A showing additional details of components of the system example of the present invention;

FIG. 2 is a schematic diagram of an example of an image displayed on a display of a fluoroscopy system;

FIG. 3 is a schematic diagram of other aspects of the example system of FIG. 1A;

FIG. 4 is a schematic diagram of an example of an X-ray exposure area of the detector with reference to the parameters of FIG. 3;

FIG. 5 is a schematic diagram of an example of a collimator according to the invention;

FIG. 6 is a schematic diagram of an example of an exposure area of the image intensifier at a certain rotation angle of the collimator of FIG. 5;

FIG. 7 is a schematic diagram of an example of a light exposure pattern of the sensor of FIG. 5 at a certain rotation angle of the collimator;

FIG. 8 is a schematic diagram of an example of a process for reading sensor pixel values;

FIG. 9 is a schematic diagram of an example of a process for reading sensor pixel values;

FIG. 10A is a schematic diagram of a top view of an example of a collimator of the present invention;

FIG. 10B is a schematic diagram of a bottom view of the example collimator of FIG. 10A;

FIG. 10C is a schematic diagram of a cross-sectional view of the example collimator of FIG. 10A;

FIG. 11A is a schematic diagram of the main components of another example collimator of the invention;

FIG. 11B is a schematic view of the portion of FIG. 11A in an operating configuration;

FIG. 11C is a schematic view of the cross-section of FIG. 11B;

FIG. 11D is a schematic diagram of components of the example collimator of FIG. 11B;

FIG. 12A is a schematic diagram of the main modules of another example collimator of the invention;

FIG. 12B is a schematic diagram of the module of FIG. 12A in an operational configuration;

FIG. 13A is a schematic diagram of another example collimator of the present invention;

FIG. 13B is a schematic diagram of another example collimator of the present invention;

FIG. 14A is a schematic diagram of the main components of another example collimator of the invention;

FIG. 14B is a schematic view of the components of FIG. 14A in an operating configuration;

FIG. 15 is a schematic diagram of another 4 examples of collimators of the present invention and the qualitative exposure of the collimators as a function of distance from the center of rotation;

FIG. 16 is a schematic diagram of another 4 collimator examples of the present invention;

FIG. 17A is a schematic diagram of an example of an ROI that is not generally located around a center of rotation;

FIG. 17B is a schematic diagram of an example of changing the rotational speed configuration of the collimator to improve the image quality of the ROI of FIG. 17A;

FIG. 18 is a schematic diagram of an example of a non-rotating collimator and its image producing effect shown on a display;

FIG. 19 is an example of the ROI of FIG. 17A and a collimator that can shift the center of rotation to the center of the ROI;

FIG. 20A is the same collimator as the example of FIG. 5, provided herein for visual comparison with the collimator of FIG. 20B;

FIG. 20B is an example of a version of the collimator of FIG. 5 with a larger diameter and a longer sector aperture for avoiding image shadowing during collimator transfer; and

FIG. 21 is a flow chart referring to FIG. 1B, illustrating a basic fluoroscopy process using the eye tracker.

Detailed Description

Referring now to FIG. 1A, a typical arrangement of a fluoroscopic clinical environment is shown.

The X-ray tube 100 produces X-ray radiation 102 which occupies a relatively large solid angle upward toward a collimator 104. The collimator 104 blocks a portion of the radiation, allowing a smaller solid angle of radiation to continue to be emitted upward, through a bed 108, typically made of a material that is relatively transparent to X-ray radiation, and through a patient 110 lying on the bed 108. A portion of the radiation is absorbed and scattered by the patient and the remaining radiation reaches the typically circular input region 112 of the image intensifier 114. The input area of the image intensifier typically has a diameter size of 300mm, but may be different for each model and technology. The image produced by the image intensifier 114 is captured by the camera 116, processed by the image processor 117, and then displayed as an image 120 on the display 118.

Although the present invention is primarily described with reference to the combination of the image intensifier 114 and camera 116, it should be understood that these elements may be replaced by digital radiography sensors of any technology, such as CCD or CMOS flat panels or other technologies such as amorphous silicon with scintillators on the plane 112. One such example is CXDI-50RF from Canon, Lake Success, New York. The term "detector" will be used to include any of these technologies, including any image intensifier in combination with any camera, and any type of flat panel sensor or any other device that converts X-rays to electrical signals.

The terms "region" and "region" are used arbitrarily in the detailed description of the invention, and they are used synonymously.

The term "X-ray source" will be used to provide a broad interpretation of a device having a point source of X-rays, and not necessarily having the shape of a tube. Although the term X-ray tube is used in the present examples in the common general nomenclature convention in the art, it is intended herein that examples of the present invention are not limited to a narrow interpretation of an X-ray tube, and any X-ray source (e.g., radioactive material configured as a point source) may be used in these examples.

An operator 122 stands near the patient to perform the treatment procedure while viewing the image 120.

A foot switch 124 is provided near the operator. When the switch is depressed, continuous X-ray radiation (or alternatively, high frequency pulsed X-rays as described below) is emitted to provide a cine image 120. The intensity of the X-ray radiation is typically a trade-off between low intensity radiation, which is desirable to reduce radiation to the patient and operator, and high intensity radiation, which is desirable to obtain a high quality image 120 (high S/N). With low intensity X-ray radiation and the resulting low radiation image intensifier input region, the S/N of image 120 may be such that image 120 becomes useless.

The coordination system 126 is a reference cartesian coordinate system where the Y-axis points to the page and the X-Y are planes parallel to the plane of, for example, the collimator 104 and the image intensifier input pad 112.

It is an object of the invention to provide a high exposure at the input area of the image intensifier at the desired ROI, thus providing a high S/N image while reducing the exposure of other parts of the image intensifier area, at the expense of a low image quality (low S/N). With this arrangement, the operator can see a clear image at the ROI and obtain an image sufficiently good for general positioning in the remaining image area. It is another object of the present invention to provide more complex image portions in an image, where each portion results from a different degree of X-ray radiation as desired for a particular application. It is a further object of the present invention to provide a different method of reading data from an image sensor. In the context of the examples provided throughout this detailed description of the invention, when the S/N of one region is compared to the S/N of another region, the S/N of pixels having the same object (e.g., patient and operator hands and tool) transmittance are compared. For example, when region a is described as having a lower S/N than region B, then it is assumed that the X-ray transmission of the subject for both regions is uniform and the same across the region. For example, in the center of region A, only 1/2 of the ray that reaches the object is sent to reach the image intensifier, the S/N of region B is compared to that of region A, which is also only 1/2 of the ray that reaches the object is sent to reach the image intensifier. S (signal) for region a is the average reading of region a (time average or region average (if enough pixels are included statistically)). S (signal) for region B is the average reading of region B (time average or region average (if enough pixels are included statistically)). Scattered radiation is not considered in the detailed description of the invention for simplicity of discussion. Means for influencing and reducing scattered radiation are well known in the art.

In the following examples, the noise statistics assumption is a gaussian distribution, which satisfies most practical aspects of practicing the invention and may well serve as a clear expression of the detailed description examples of the invention. This is not a limitation of the present invention and, if desired, the mathematical analysis associated with gaussian statistics may be replaced with poisson statistics (or other statistics) without reducing the scope of the present invention. The noise value associated with each signal is represented by the standard deviation of the poisson statistics for that signal, known in the art as poisson noise.

In addition, the DPP per pixel dose is described throughout this specification in the same sense, i.e. when the DPP of pixel A is compared to the DPP of pixel B, it is assumed that the object transmission for both pixels is the same.

A more detailed arrangement example of a fluoroscopic clinical environment according to the present invention is shown in fig. 1B and fig. 21. The operator 122 depresses the foot switch 124 to activate the X-ray (step 2724). An eye tracker 128 (e.g., Eyelink 1000 from SR research ltd, catata, ontario, canada) or any optional input device provides an indication of where the operator 122 is looking (step 2728). This information is typically provided with respect to the display 118. This information, the "point of regard", may be provided, for example, in (X, Z) coordinate form, in the plane of the display 118, using a coordinate system 126. It will be appreciated that in this example, the plane of the display 118, and thus the image 120, is parallel to the (X, Z) plane of the coordinate system 126. Other coordinate systems are also available, including a coordinate system that is bound to the display 118 and rotates with the display 118 as the display 118 rotates relative to the coordinate system 126.

The data input by input 128 is provided to controller 127, which is essentially a computer, such as any PC computer. If the controller 127 determines that the operator's gaze is not fixed on the image 120, the X-ray tube 100 is not activated (step 2700). Otherwise, at step 2710, the X-ray tube 100 is activated and emits X-ray radiation to the collimator 104.

An example of an image 120 displayed on the display 118 is now described with reference to FIG. 2. In this example, the dashed circle line 204 indicates the boundary between the portion 200 of the image and the portion 202 of the image, which constitute the complete image 120. In this example, it is desirable to obtain good image quality in portion 200, meaning that the X-ray DPP is higher for portion 200, and to have lower image quality in portion 202 is acceptable, meaning that the DPP of portion 202 is lower.

It should be understood that the two sections 200 and 202 are provided herein only as examples of embodiments of the present invention, which are not limited to the present example, and that the image 120 may be divided into any number of sections by controlling the shape of the aperture in the collimator and the pattern of collimator movement. Examples of these are provided below.

It should be understood that DPP should be interpreted as the X-ray dose delivered to a portion representing a pixel of image 120 that produces a pixel readout (excluding absorption by the patient or other elements not part of the system, such as the operator's hands and tools) used to construct image 120.

Referring now to fig. 3, a typical collimator 104 with a circular aperture 304 is introduced into the X-ray path such that only X-rays 106 emanating from the focal point 306 of the X-ray tube 100 and passing through the aperture 304 reach the circular input surface 112 of the image intensifier 114, while other X-rays 102 are blocked by the collimator. This arrangement exposes the entire input area 112 of the image intensifier to the same DPP. This arrangement does not provide the functionality of one DPP of section 300 associated with section 200 of fig. 2 and another DPP of section 302 associated with section 202 of fig. 2. The diameter of the input area 112 is B as shown in fig. 3.

D1 represents the distance from the X-ray focal point 306 to the aperture 104. D2 represents the distance from the X-ray focal point 306 to the image intensifier input surface 112.

Reference is now made to FIG. 4, which defines part of the present example of an image intensifier input surface 112 to support the examples of the present invention. In this example, the portion 300 is a circular region of diameter R1 centered on the image intensifier circular input region 112. Portion 302 has an annular shape with an inner diameter R1 and an outer diameter R2. R2 is also typically the diameter of the image intensifier input area.

Referring now to FIG. 5, one embodiment of a collimator for providing one DPP for section 300 and another DPP for section 302 is provided.

Collimator 500 is constructed substantially as a circular flat plate of X-ray absorbing material (e.g., lead, typically 1-4mm thick) having a diameter greater than r 2. The aperture 502 of the collimator 500 is configured as a circular notch 504 with radius r1 at the center of the collimator and a fan-shaped notch 506 with radius r2 and angle 508. It should be understood that the term sector is used to indicate both a sector of a circular area and a sector of an annular area, as described above and below.

In this example, R1 and R2 of aperture 502 are designed to provide R1 and R2 of fig. 4. When collimator 500 is in the position of collimator 104 of FIG. 4, r1 and r2 may be calculated using the following equation:

r1＝R1/(D2/D1)

r2＝R2/(D2/D1)

in this example, the angular span 508 is 36 degrees, i.e., 1/10 degrees of circumference. Collimator 500 may be rotated about its center as indicated by arrow 512. A weight/weight 510 may be added to balance the collimator 500 and to ensure that the coordinates of the center of weight coincide with the coordinates of the center of rotation in the collimator plane, thus avoiding system vibrations that may result from collimator imbalance. After completion of one 360 degree rotation, the DPP of portion 302 is 1/10 of the DPP of portion 300.

It should be appreciated that angle 508 may be designed to meet any DPP ratio requirement. For example, if angle 508 is designed to be 18 degrees, after one full rotation of aperture 500, the DPP of portion 302 is 1/20 of the DPP of portion 300. The discussion of the current example will proceed with angle 508 being 36 degrees.

After one rotation of collimator 500 is complete, camera 116 captures one frame of data gathered by the sensors during one full rotation time of collimator 500, this frame including data values read from the pixel set of the camera sensors. This will now be explained in more detail, and a camera based on a CCD (charge coupled device) sensor is provided as an example, such as a TH 8730CCD camera from THALES electronics DEVICES of Velizy Cedex, france.

In this example, synchronization of the rotation of the camera 116 and the collimator 500 is performed using a protrusion 514 constructed on the collimator 500 by a light sensor 516 (e.g., EE-SX3070 from OMRON Management Center of America, Schaumburg, Ill.).

When an interrupt signal for the protrusion 514 is received from the light sensor 516, the lines of the camera 116 sensors are transferred to their shift registers and the pixels begin a new integration cycle. The data of the previous integration period is read from the camera. When the protrusion 514 interrupts the light sensor 516 again, the accumulated signal is again transferred to the shift register of the camera sensor 116, forming the next frame to be read out.

In this way, one frame is generated for each complete collimator cycle. For each frame, the DPP in portion 202 of image 120 is 1/10 of the DPP in portion 200 of image 120.

To provide additional views of the above, reference is now made to FIG. 6, which depicts an exposure image of image intensifier input 112 for the instantaneous position of rotary collimator 500. In this position, the circular area 600 and the sector area 602 are exposed to radiation, while the remaining sector area 604 is blocked from being exposed to radiation by the collimator 500. As collimator 500 rotates, sectors 602 and 604 rotate with it, while circular area 600 remains unchanged. During one cycle of constant speed rotation of collimator 500, each pixel outside of region 600 is X-rayed at 1/10 times that pixel in region 600 is X-rayed, so the DPP received by the pixel outside of region 600 is 1/10 of the DPP received by the pixel in region 600.

In fig. 7, an equivalent optical image projected on a camera sensor 710 is shown, where region 700 of fig. 7 is equivalent to region 600 of fig. 6 and region 702 of fig. 7 is equivalent to region 602 of fig. 6. The output image projected by the image intensifier onto the sensor 710 is represented by the numeral 712. 714 is a typical sensor area that is outside the range of the image output of the image intensifier.

For each frame, in addition to compensating for the linear response characteristic of each pixel with typical offset and gain corrections, the signals of the pixels of portion 202 need to be multiplied by a factor of 10 to produce image 120, so that the brightness and contrast display of portion 202 is similar to that of portion 200. This method, described herein with reference to a particular example, will be referred to as pixel "normalization". The normalization scheme (i.e. the shape, velocity and position of the collimator) is made according to the X-ray exposure scheme.

To produce a motion picture of 10 frames per second (fps), the collimator 500 needs to be rotated at a speed of 10 turns per second (rps). To produce a 16fps film, the collimator 500 needs to be rotated at a speed of 16 rps.

With each 360 degree rotation, a complete exposure of the input area 112 is completed. The exposure period (EC) is thus defined as the minimum number of rotations of the collimator 500 for providing the minimum complete design exposure of the input area 112. In the collimator 500 example of fig. 5, the EC needs to rotate 360 degrees. For other collimator designs, such as the one in fig. 13A, EC needs to be rotated 180 degrees, and for the one in fig. 13B, EC needs to be rotated 120 degrees.

It should be appreciated that examples of collimators, X-rays projected on the image intensifier input area 112, images projected on the camera sensor (or flat panel sensor), and images displayed on the display 118 are illustrated in a general manner that omits possible geometric issues, such as image inversion due to a lens image that may be different if mirrors are also used, or rotation directions that are shown clockwise throughout the description but may be different depending on the particular design and orientation of the viewer. It should be understood that those skilled in the art understand these options and have the correct explanation for any particular system design. It should be understood that the camera frame reading scheme described above with reference to collimator 500 may be different:

1. the reading of the frame does not have to be at the moment when the protrusion 514 interrupts the light sensor 516. Rather, this may be done at any stage of collimator 500 rotation, as long as each EC is done at the same stage.

2. More than one frame is read during one EC. However, it is desirable to read an integer number of frames for each EC. As such, the read frame includes the complete data of one EC, making it easier to establish one display frame presented on the display 118 in multiple ways:

a. the pixel values of all frames of an EC are summed to produce a complete exposure image. The pixel values of all frames of the next EC are then summed to produce the next complete exposure image. Thus, each time EC is completed, the picture on the display is replaced by a temporally continuous image. The normalization of the pixel values may be done independently for each frame, or may be done only once for the sum of the frames, or any combination of the frames.

b. For the example of the method, assume that the camera provides 8 frames during one EC. In this example, all 8 frames with sequence numbers from 1 to 8 are stored in the frame memory, while the first display frame is generated from these frames as described above (summing all frames and normalizing the pixel values). The resulting image is then displayed on the display 118. When frame 9 is available (after 1/8 EC), frame 1 is replaced with frame 9 in the frame memory and frames 9, 2, 3, 4, 5, 6, 7, 8 are processed (summed, normalized) to produce a second display frame that can now be displayed on display 118 after 1/8 EC. After another 1/8EC, frame 10 is obtained and stored in the location of frame 2. Frames 9, 10, 3, 4, 5, 6, 7, 8 are processed to produce a third display frame. In this manner, the sequence of film images is displayed to the user on the display 118 using a frame memory managed by a FIFO (first in first out) method and generating a display frame with each new frame obtained from the sensor.

c. In another embodiment of the invention, only the frame pixels of the pixels exposed by the X-rays according to the rules of collimator shape and action are summed during the integration time of the acquired frames. This may be 1/8 for EC time in example b above. The pixels to be summed to produce the image are (1) from region 700 and (2) in a sector of 2x order of angle (angular span 508 of collimator sector 506). The reason for 2X is that in 1/8 at integration time, the collimator is rotated 1/8 of EC. A sector angle (angle 508) slightly greater than 2 · is desired to compensate for the accuracy limitation. This summing method significantly reduces the number of pixels involved in the summing process, thus reducing computation time and computational resources.

d. In another embodiment of the present invention, the pixel processing is limited to the pixels specified in c above. This approach significantly reduces the number of pixels involved in the summation process, thus reducing computation time and computational resources.

e. In another embodiment of the present invention, the stored pixels are limited to the pixels specified in c above. This storage method significantly reduces the number of pixels included in the memory, thus reducing the storage requirements.

f. In another embodiment of the invention, any of the methods described in this section (a-general concept, b-as specific examples of a, c, d, and e) may be combined into implementations using any combination of some of these methods.

3. One frame is read during more than one EC. In another embodiment, the collimator may be operated to provide an integer number of ECs for each frame received from the sensor. For example, after 2 EC of the collimator, one frame is read from the sensor. After normalizing the pixel values of this frame, it may be displayed on the display 118.

It should be understood that in many designs, the frame rate provided by the sensor is dictated by the sensor and associated electronics and firmware. In this case, the rotation speed of the collimator 500 may be adapted to the sensor characteristics such that the time of one EC is the same as the time of receiving an integer number of frames (one frame or more) from the sensor. It is also possible to set the rotation speed of the collimator such that an integer number of ECs are done during the time period in which the frame is obtained from the sensor.

The above description of reading of frames is particularly sufficient for a CCD type sensor, whether the CCD camera is mounted to the image intensifier or a flat panel sensor located approximately at the plane 112 of fig. 3 in place of the image intensifier and camera. A special feature of CCDs is to acquire the values of the entire frame, i.e. all pixels of the sensor, at one time. This is followed by sequential transmission of the analog values to an analog-to-digital converter (a/D). Other sensors, such as CMOS photosensors, typically read frame pixels one after another, which is known as the rolling shutter method. The method of reading the sensor frame synchronized with the collimator EC is applicable to such a sensor regardless of the frame reading method used. The "random access" function of reading the pixels of a sensor (e.g., a CMOS sensor) provides another embodiment of the present invention. Unlike a CCD sensor, the order in which the pixels are read from the CMOS sensor can be any order desired by the system designer. The following embodiments use this function. In this context, a CMOS sensor represents any sensor that supports pixel readout in any order.

Reference is now made to fig. 8. The embodiment of fig. 8 is also illustrated using the example of an image intensifier and CMOS camera, but it will be appreciated that the method of this embodiment can also be used for flat panel sensors and other sensors capable of random access pixel reading.

The output image of the image intensifier 114 is projected onto a region 712 of the sensor 710. Depending on the instantaneous position of the rotating collimator 500, the circle 700 and the sector 702 are instantaneously illuminated, and the sector 704 and the sector 714 are not illuminated, together with the position of the collimator 500. Sectors 702 and 704 rotate with the rotation of collimator 500 as indicated by arrow 706.

For purposes of this example, pixels that precede a radial line, such as 702A or 800A, are pixels whose centers are on the radial line or clockwise of the radial line. The pixels behind the radial line are the pixels whose centers are in the counterclockwise direction of the radial line. Sector 702 includes, for example, pixels after radial line 702A and pixels before radial line 702B. For example, in an embodiment mode in which a frame is read from the sensor once in EC, pixels adjacent to radial line 702A have just begun exposure to the output image of the image intensifier, and pixels adjacent to radial line 702B have just finished exposure to the output image of the image intensifier. The pixels in sector 702 are partially exposed at each location where they are located between 702A and 702B. In this example, the pixels in the sector between radial lines 702B and 800B have not been read after exposure to the image intensifier output.

In the present example of the present embodiment, the instantaneous angular position of the radial line 702A is K.360 degrees (K times 360, K being an integer indicating the number of EC from the start of rotation). In the example of collimator 500, the angular span of sector 702 is 36 degrees. Thus, radial line 702B is at an angle of K360-36 degrees. At this position of the collimator, the pixel reading period of sector 800 begins. Radial line 800A is defined to ensure that all pixels following this radial line are fully exposed. This angle can be determined using R1 of fig. 5 and the size of the pixel projected onto fig. 5. To calculate the theoretical minimum angular span between 702B and 800A to ensure that pixels adjacent to 800A are also fully exposed, the arc of radius R1 should be considered to have a chord length in length of 1/2 pixel diagonal. This determines the minimum angular span between 702B and 800A to ensure that all pixels in sector 800 are fully exposed. In a more specific implementation, assuming that region 712 has approximately 1000 pixels in the vertical direction and 1000 pixels in the horizontal direction, and that R1 is approximately 1/4-1/2 of R2 (see fig. 4), and considering the tolerances of this design and implementation, the useful arc length of radius R1 should be, for example, the length of a 5 pixel diagonal. This means that the angular span between 702B and 800A will be about 2.5 degrees. That is, at the instant of the example of FIG. 8, the angular position of radial line 800A is K.360- (36+2.5) degrees.

In this particular example of the present embodiment, the angular span of the sector 800 is also selected to be 36 degrees. Thus, at the instant in the example of FIG. 8, the angular position of ray 800B is K360- (36+2.5+36) degrees.

In fig. 8, the angular span of sector 800 is drawn to illustrate a smaller angle than the angular span of sector 702 to emphasize that the angles need not be the same and that the example provided here is the same only for the purpose of being one specific example of this embodiment.

Having determined the geometry of sector 800, the pixels of that sector are now read from the camera sensor. In a typical CMOS sensor, each pixel is reset after it is read so that the pixel can accumulate signals again from zero. In another embodiment, all pixels of sector 800 are read during the first phase and pixels are reset during the second phase. The read and reset cycle of sector 800 must end within the time that sector 702 is rotated an angular distance equal to the angular span of sector 800 so that the system is ready in time to read the next sector that is the same angular span as sector 800, rotated clockwise by the amount of the angular span of sector 800 relative to the angular position of sector 800. In this example: 36 degrees.

In the example described above, when the collimator 500 is rotated at 10rps, a 36 degree span sector 800 will have 10 orientations in one EC, the orientations being 36 degrees apart, and the speed of the pixel read and reset cycle being 10cps (cycles per second).

It should be understood that this embodiment may be implemented in different specific designs.

For example, the angular span of sector 800 may be designed to be 18 degrees while the angular span of sector 702 is still 36 degrees and the collimator 500 rotates at 10 rps.

In this example, the sector 800 would have 20 orientations in one EC, 18 degrees apart, with a speed of 20cps (cycles per second) for the pixel read and reset cycles.

In another embodiment, dark noise accumulated by pixels in sector 704 after radial line 800B and before radial line 802A is cancelled by another reset period of pixels located in sector 802 (after radial line 802A and before radial line 802B). This reset process ideally occurs in sector 802 particularly close to and before sector 702. The resetting of all pixels of sector 802 must be completed before the radial line 702A of the rotated sector 702 reaches the pixels of sector 802. Otherwise, the angular span and angular position of the reset sector 802 are designed in a similar way and with similar considerations as those used to determine the sector 800.

The pixels read from sector 800 should be normalized and can be used to generate display frames in a similar manner to those described in section 2 "read more than one frame during one EC", above, where in the current embodiment only sector pixels are read, stored and processed, rather than the entire sensor frame.

In this embodiment, after the pixels of the last read sector are normalized, the processed pixels can be used to directly replace the corresponding pixels in the display frame. In this way, the display frame is refreshed in a similar pattern to the radar beam sweep, each time the next sector of the image is refreshed. After the 360/(angular span of read sectors) refresh, the entire display frame is refreshed. This provides a simple picture refresh scheme.

Attention is now directed to fig. 9. Unlike fig. 8, where the read sector includes a complete set of pixels after radial line 800A and before radial line 800B, in the present invention, the read region geometry is divided into two parts: a circular area 700 and a sector 900. The sector 900 of the embodiment of fig. 9 contains pixels that are after radial line 900A and before radial line 900B, and also after radius R-1 and before R-2. In this example, the pixels before the radius are pixels having a distance to the center less than or equal to the radius R, and the pixels after the radius R are pixels having a distance to the center greater than R. The pixels of region 700 are all those pixels located before R-1.

In this embodiment, the same reading and processing of the pixels of portion 900 is performed in the same manner as described with reference to the embodiment of FIG. 8, and is also applicable to the reset sector 802.

The pixels of region 700 are treated differently.

In one implementation of the current embodiment, the pixels of area 700 may be read one or more times during an EC period and processed as described above for reading the entire CMOS sensor, or area 700 may be read once during one or more ECs and processed as described above for reading the entire CMOS sensor.

It will be appreciated that for each reading method, a normalization process of the pixels must be performed to obtain a display frame in which all pixel values represent exposures of the same sensitivity.

Attention is now directed to fig. 10, which provides one example of the collimator of the present invention in combination with a motion system that provides a rotation function to collimator 500.

Fig. 10A is a top view of the collimator and rotation system of the present example.

Fig. 10B is a bottom view of the collimator and rotation system of the present example.

Fig. 10C is a cross-sectional view of a-a of fig. 10A.

Fig. 10A shows a collimator 500 and an aperture 502 (with other details removed for clarity). Pulley 1000 is mounted on top of collimator 500 in a position concentric with the collimator. The pulley 1002 is fitted to a motor 1012 (see the motor in fig. 10B and 10C). A belt 1004 couples pulley 1000 with pulley 1002 to transmit rotation of pulley 1002 to pulley 1000, thus providing the desired rotation of collimator 500. Belt and pulley system examples 1000, 1002, and 1004 present a flat belt system, but it should be understood that any other belt system may be used, including circular belts, V-belts, multi-grooved belts, ribbed belts, film belts, and time-counting belt systems.

Fig. 10B shows the bottom of fig. 10A, showing further components not previously shown. A V-shaped circular track 1006 is shown coaxial with collimator 500 (see cross-section a-a of 1006 in fig. 10C). Three wheels 1008, 1010, and 1012 are in contact with the V-groove track 1006. The axis of rotation of the 3 wheels is fitted to an annular stationary part 1016 (not shown in fig. 10B) of the relative body fixed to the X-ray tube. This configuration provides support for the collimator 500 to be in a desired position relative to the X-ray tube (e.g., the position of the collimator 104 of fig. 3), while at the same time providing 3 wheels 1008, 1010, and 1012 and rails 1006 for rotation of the collimator as desired.

Rotation of motor 1014 is transmitted by pulley 1002 to collimator 500 through belt 1004 and pulley 1006. The collimator is then supported for rotation by rails 1006 that slide on wheels 1008, 1010, and 1012.

It should be understood that the rotation mechanism described herein is only one example of a possible implementation of a rotation mechanism for rotating a collimator. The rotation mechanism may instead use any type of gear transmission mechanism including straight, helical, bevel, hypoid, crown and screw gears. The rotation mechanism may use a high friction surface cylinder for 1002 and bring 1002 into direct contact with the edge of the collimator, so that the belt 1004 and pulley 1000 are not required. In another implementation, the collimator 500 may also be configured as a rotor of a motor, around which a stator is additionally arranged.

In the illustration of the collimator of fig. 5, the protrusion 514 and the light sensor 516 are represented as elements that provide tracking of the angular position of the collimator 500 for synchronization between the collimator angular position and the sensor reading process. These elements are presented as an implementation example. The implementation means for tracking the rotational position may be implemented in a number of other ways. In the example of FIG. 10, the motor 1002 has an attached encoder, such as from Maxon precision Motors, Fall River, Mass. A simple encoder can be constructed by recording black and white binary code strips on the circumference of the collimator 500 and reading the strips using a light sensor, such as a Newark (http:// www.newark.com) TCRT5000 reflective light sensor.

The collimator described above has a fixed aperture and cannot be changed after the collimator has been manufactured.

It should be understood that in other embodiments of the present invention, the mechanical design of the collimator assembly may be made to accommodate replaceable collimators. In this way, different apertures can be fitted to the collimator assembly according to the requirements of a particular application.

In other examples of practice of the invention, the collimator may be designed to have a variable aperture within the collimator assembly. This is shown in figure 11.

The collimator of fig. 11 is constructed of two superimposed collimators as shown in fig. 11A. One collimator is 1100, which has an aperture 1104 and a counterweight 510, such that the center of gravity of this collimator is the center of rotation of the collimator. The second collimator is 1102 with an aperture 1105 and a counterweight 511 such that the center of gravity of this collimator is the center of rotation of the collimator. In both collimators, the aperture geometry is a combination of a central circular hole of radius r1 and a fan-shaped hole of radius r2 and a sector angular span of 180 degrees. In practice, collimator 1102 is of the same design as collimator 1100, but flipped upside down.

Placing collimators 1100 and 1102 one above the other and coaxially as shown in fig. 11B results in a combined aperture which is the same as the aperture in the collimator of fig. 5. By rotating collimator 1100 relative to collimator 1102, the angular span of sector 508 may be increased or decreased. In this example, the angular span of the sector 508 may be set in the range of 0-180 degrees. In this example, ring 1108 holds collimators 1100 and 1102 together, as shown in FIG. 11C, which is a cross-sectional view B-B of FIG. 11B.

Reference is now made to fig. 11C (weights 510 and 511 are not shown in this cross-sectional view). In this example of the invention, ring 1108 is shown holding collimators 1100 and 1102 together, allowing them to be rotated one relative to the other to set angular span 508 of sector 506 as desired. An example of a locking mechanism for maintaining collimators 1100 and 1102 at a relatively desired angle is shown in FIG. 11D. In fig. 11D, ring 1108 does not show collimators 1100 and 1102 for clarity. Portion 1110 is cut away in the figure to expose the U-shape 1112 of ring 1108, in which collimators 1100 and 1102 are received. After the desired angular span 508 has been set, screws 1114 fitted into threaded holes 1116 are used to lock collimators 1100 and 1102 in place. To change the angular span 508, the operator may release the screws 1114, reorient the collimators 1100 and/or 1102, and then tighten the screws 1114 again to set the collimator positions.

The example of FIG. 11, including manual adjustment of the angular span 508, is provided as an example of an implementation of the present invention. Many other options are available. Fig. 12 shows another example. In this example, the angular span 508 may be computer controlled. The mechanism of fig. 12 is primarily a structure including two units similar to the unit of fig. 10, but with a small change including the elimination of the pulley 1000 instead of using the edge of the collimator as a pulley. For clarity, the balancing weights 510 and 511 are not shown in the drawings.

In fig. 12A, the bottom unit including collimator 500 is essentially the assembly of fig. 10, but with pulley 1000 removed and instead using the edge of collimator 500 as a pulley. In the top unit that includes collimator 1200, the bottom assembly is identical to the bottom assembly when rotated 180 degrees about an axis perpendicular to the page, unless motor 1214 is rotated 180 degrees further so that it is below the pulley, similar to motor 1014. This is not required for this example, but in some design cases, it helps to keep the space above the assembly of fig. 12 free of unwanted items. Fig. 12B shows these 2 components together so that collimators 500 and 1200 are close to each other and concentric. In the assembly of fig. 12B, each of collimators 500 and 1200 may be rotated independently. For each collimator, the angular position may be known by any encoding system, including the examples provided above.

In one example of using the assembly of FIG. 12B, angular span 508 is not set when collimator 500 is stationary and collimator 1200 is rotated until the desired angle 508 is reached. Then, both collimators are rotated at the same speed to provide the X-ray exposure pattern example as shown above. It should be understood that it is not necessary to stop any of the collimators to adjust the angle 508. Conversely, during rotation of the two collimators, the rotational speed of one collimator relative to the other may be changed until the desired angle 508 is reached, and then the two collimators continue to rotate at the same speed.

It should be appreciated that mechanisms having the functionality of the example shown in FIG. 12B, for example, can be used to introduce more complex exposure patterns. With this mechanism, the angle 508 may be changed during an EC to produce multiple exposure modes. For example, the angle 508 may increase in the first half of the EC and decrease in the second half of the EC. This will result in an exposure pattern of 3 different exposures (it will be appreciated that the boundaries of the areas exposed by the sectors 506 are not sharp and that the width of these boundaries depends on the angle 508 and the speed at which this angle is changed relative to the rotational speed of the collimator).

It should also be understood that any collimator of the present invention may rotate at variable speeds at EC and affect the geometry of the exposure. For example, the collimator 500 of fig. 5 may be rotated at one speed for the first 180 degrees of EC and at twice the speed for the other 180 degrees of EC. In this example, the DPP of the area exposed by sector 506 during the first half of the EC is twice the DPP of the area exposed by sector 506 during the second half of the EC, the boundary DPP between the two halves being changed stepwise. The central area exposed through the circular aperture 504 will have a 3 rd order DPP. Other rotational speed configurations may produce other exposure shapes. For example, 3 different rotation speeds at 3 different parts of the EC will produce 4 regions with different DPP.

The examples provided above propose collimators having apertures with a similar basic shape comprising a central circular opening in combination with a fan-shaped opening. These examples are used to illustrate many aspects of the present invention, but the present invention is not limited to these examples.

Referring now to fig. 13A, another example of an aperture of the present invention is shown. In this example, the aperture of the collimator 1300 is made up of a circular hole 1302 concentric with the collimator edge, a scalloped hole 1304, and a scalloped hole 1306 located in the opposite direction of 1304 (two scallops spaced 180 degrees apart). If it is desired that, for example, the DPP exposed to the annular region of fig. 6 (including sectors 602 and 604) is 1/10 of the DPP of region 600 of fig. 6, then each of sectors 1304 and 1306 may be set to 18 degrees and collimator 1300 may then be rotated only 180 degrees relative to the 360 degrees required by the collimator of fig. 5 to complete an EC. Furthermore, for 10fps, the rotation speed of the collimator 1300 should be 5rps, instead of 10rps in the case of the collimator 500 of fig. 5. In addition, for example, the balancing weight 510 of fig. 5 is not needed for the collimator 1300 of fig. 13A, because it is balanced by its own shape.

Another example of a collimator according to the present invention is provided in fig. 13B. The aperture of collimator 1310 is made up of circular hole 1312, fan 1314, fan 1316, and fan 1318 concentric with the collimator edge, the three fans being 120 degrees apart. If it is desired that, for example, the DPP exposed to the annular region of fig. 6 (including sectors 602 and 604) is 1/10 of the DPP of region 600 of fig. 6, then each of sectors 1314, 1316, and 1318 may be set to 12 degrees, and then collimator 1310 may be rotated only 120 degrees relative to the 60 degrees required by the collimator of fig. 5 to complete an EC. Furthermore, for 10fps, the rotation speed of the collimator 1300 should be 10/3rps, instead of 10rps in the case of the collimator 500 of fig. 5. Also, for example, the balancing weight 510 of fig. 5 is not needed for the collimator 1310 of fig. 13B because it is balanced by its own shape.

It should be appreciated that the relationships and methods for rotating the collimator in the example of fig. 13A and 13B, and reading pixel values from the light sensors as described above in connection with the collimator example of fig. 5, can be fully migrated into the collimator example of fig. 13A and 13B by adjustments apparent to those skilled in the art. For example, the collimator of fig. 13B and the CMOS camera pixel read sensor 800 of fig. 8 can be supplemented by adding two pixel read sensors, each in combination with one of the 2 additional aperture sectors of fig. 13B.

Some of these changes and comparisons are shown in the following table, representing characteristic and implementation examples of differences between 3 different collimator examples.

Collimator	FIG. 5	FIG. 13A	FIG. 13B	Note
					Central circular aperture	Is that	Is that	Is that
Number of aperture sectors	1	2	3
					Angular span of sector	36 degree	18 degree	12 degree	For a 1:10DPP ratio
Angular interval of sector	NA	180 degrees	120 degrees
					EC rotation	360 degree	180 degrees	120 degrees
rps	10	5	10/3	For 10rps
					Fps of 10rps	10	20	30

Figures 11 and 12 provide one example of how the collimator of figure 5 can be implemented in the form of a variable angular span 508 of sectors 506.

Fig. 14 provides an example of how the collimator of fig. 13A may be configured so that the angular span of sectors 1304 and 1306 may be adjusted as desired.

An example of 2 collimators 1400 and 1402 is shown in fig. 14A. The grey background rectangle blocks are used to provide a better display of the collimator solid area and the aperture holes, but are not part of the structure.

This is also true of fig. 14B. Each collimator has an aperture consisting of an annular hole concentric with the edge of the collimator and two sector holes, each sector hole having an angular span of 90 degrees, the sectors being spaced 180 degrees apart. When collimators 1400 and 1402 are placed one above the other and concentric, the combined collimator of fig. 14B is provided. The size and shape of the aperture of the collimator in fig. 14B is the same as the size and shape of the aperture of the collimator in fig. 13A. However, in the case of the combination shown in FIG. 14B, the angular span of aperture sectors 1404 and 1406 can be varied by resetting the ratio of collimators 1400 and 1402 relative to one another. This may be accomplished using any of the methods described above with reference to fig. 11 and 12.

It should be understood that similar designs may provide variable angular spans for the aperture sectors and other aperture designs of the collimator 1310 of fig. 13B.

In the aperture design described above, the aperture shape is designed to provide two regions with two different DPPs at a constant rotational speed.

FIG. 15A shows such a collimator and shows a graphical representation of the exposure for two levels of DPP for different distances r from the center. Other apertures may be designed to provide any desired exposure profile. Some examples are shown in fig. 15B, 15C, and 15D. All collimators of fig. 15 have an aperture design rotated 360 degrees for one EC.

The aperture features in the collimator of fig. 15 may be combined with the aperture features in the collimator of fig. 13. An example of such a combination is shown in fig. 16, which shows 4 collimators with 4 different aperture designs. In fig. 16A, the left and right halves of the aperture are asymmetric and one EC needs to be rotated 360 degrees. Fig. 16B provides a collimator that may have an aperture that provides an exposure profile similar to (but not identical to) that of fig. 15C, but with one EC rotated only 90 degrees. Fig. 16C provides a collimator that may have an aperture that provides an exposure profile similar to (but not identical to) that of fig. 15D, but one EC includes only 360/8-45 degree rotations. Fig. 16D provides a collimator with an aperture that provides an exposure profile similar to (but not identical to) that of fig. 15D, but one EC includes only 180 degrees of rotation.

From these examples, it should be appreciated that the present invention may be implemented in a variety of designs and is not limited to the specific designs provided as examples above.

Pixel correction:

as described above, each collimator design and pixel with a different DPP is normalized to provide a suitable display frame. The normalization scheme is performed according to the X-ray exposure pattern (i.e., the shape, velocity, and position of the collimator). This normalization can be done on the basis of theoretical parameters. For example, referring to figures 7 and 5, with the collimator 500 rotating at a constant speed, the pixels of the annular composite sectors 702 and 704 receive 1/10 of the dose of the circular area 700 (in this example, the angular span 508 of sector 506 is 36 degrees). To simplify the present example, assume that one frame is read from the sensor at the end of each EC (i.e., the collimator 500 completes a 360 degree rotation). It is also assumed that all sensor pixels have the same response to the image intensifier output, whereas the image intensifier has a uniform response, and that the X-rays from the X-ray tube are uniform. The only source of built-in (i.e., system-level) differences between pixels comes from the collimator and the way it operates. In this example, the normalization based on the system design would be pixel multiplication by one or two factors to compensate for the DPP difference.

In one normalization example, the pixel values of the annular composite sectors 702 and 704 may be multiplied by 10. In another normalization example, the pixel values of the circular area 700 may be multiplied by 1/10. In another normalization example, the pixel values of the annular composite sectors 702 and 704 may be multiplied by 5, while the pixel values of the circular area 700 may be multiplied by 1/2. It is to be understood that in the description, explanation and examples of the present invention, multiplication and division are equivalent, that is, an expression similar to "multiplied by 1/10" is entirely equivalent to an expression of "divided by 10", which also means, whenever a multiplication by a value is mentioned, an optional division by its reciprocal value, and vice versa. This also applies to the multiply and divide signs used in the formula. For example, a/B also represents a · C, wherein C ═ 1/B.

The above example is relatively simple because the normalization scheme combines 2 known regions with two known DPPs. The situation may become more complex for different collimators or collimator movement schemes.

In the following example, the rotation of the collimator 500 introduces a change. That is, a variable rotation speed is used instead of a constant rotation speed, as proposed for one EC (in case of a collimator 500: 360 degrees) in the following table:

sector number	EC range (degree)	Angular rotation state
			1	0-150	Constant velocity 1
2	150-180	Constant positive acceleration
			3	180-330	Constant speed 2
4	330-360	Constant negative acceleration

This rotation pattern, in combination with the convolution of the image pixels, particularly the accelerating sectors, makes it more difficult to estimate the normalization factor.

In the collimator examples of fig. 15C and 15D, multiple "pixel rings" (pixels at a fixed distance from the center) require a suitable normalization factor. System manufacturing tolerances not included in the theoretical estimate of the normalization factor may lead to errors that will appear as ring patterns in the image displayed on the display 118.

The following calibration method provides a theoretical estimate of the unneeded factor and compensates for the calibration of manufacturing tolerances.

In this example, any collimator of the present invention may be used, as well as any rotational mode that is fixed for each EC.

The fluoroscopy system is arranged to include all fixed components related to the imaging process (X-ray tube, desired X-ray operation mode, i.e. voltage and current, possible X-ray filters, collimator, patient bed, image intensifier, camera), but not any variable parts (patient, operator hands and tools).

According to this calibration method, the required collimator is rotated in the desired pattern. A set of original frames is obtained (using any of the methods described above). The original frame is a frame derived from one or more integer numbers EC of all pixels of region 712 (fig. 7) without any manipulation of the pixels. The number of resulting original frames should be sufficient to obtain a relatively good S/N on average of the resulting original frames. It is usually sufficient to have an average original frame that is 10 times higher than the S/N of the original frame, and this can be obtained by taking an average of 100 original frames. It should be understood that more or fewer original frames may be used depending on the desired quality of the normalized frame.

One average raw frame is created when the X-rays are off and the other is created when the X-rays are on. For this example, assume that the luminance value of each pixel for display purposes ranges from zero to 255. Theoretical noiseless frames displayed in the range of 5-250 are also selected (the darkest noiseless pixels will be displayed at value 5 and the brightest exposed noiseless pixels at value 250. this causes noise bringing the pixel values into the ranges 0-4 and 251-255 to contribute their statistical display to the displayed frame).

The correction Pij (j is the frame number index in this example) for each pixel i of the original frame j is calculated using the value Ai of the pixel of the average original frame obtained with the X-ray turned on and the value Bi of the pixel of the average original frame obtained with the X-ray turned off to produce a corrected pixel Dij as follows:

(formula 1) Dij ═ di (Pij-Bi) · (245/Ai) +5

In another simpler approach, the correction may ignore the noisy visual aspects at dark and light levels, and simply correct the display range 0-255 as shown below:

(equation 2) Dij ═ (Pij-Bi) · (255/Ai)

It will be appreciated that the proposed correction is linear and it works best for systems where the image intensifier and camera have relatively linear responses.

For systems with non-linear response, more complex correction schemes, such as bilinear correction, may be used. In this example, the range of pixel values is roughly divided into two ranges. The current of the X-rays may be reduced, for example 1/2 for its normal mode of operation, such that DPP is reduced 1/2. It will be appreciated that the reduced current level depends on the nature of the non-linearity, and that an optimal bilinear correction may require an X-ray current other than 1/2. It should also be understood that DPP may also be reduced in other ways, for example by placing an aluminum plate after the collimator.

In this example, another set of raw frames is obtained using an X-ray current of 1/2. It will be appreciated that for certain applications, the S/N of these raw frames is lower than the S/N of the raw frames for standard X-ray currents. This can be compensated by using more raw frames, e.g., 200 raw frames, to generate an average raw frame of 1/2X-ray current. Suppose Mi represents the average original frame pixel values obtained when the X-ray radiation of 1/2 was turned on.

The correction example of equation 2 in this example is implemented as follows:

less than or equal to 127 for Pij

(equation 3) Dij ═ (Pij-Bi) · (127/Mi)

For Pij greater than 127

(equation 4) Dij ═ (Pij-Bi) · (255/Ai)

It should be understood that the X-ray current of Mi can be set to different levels (e.g., 1/4 which is a standard current for a particular application) and the formula is in the form:

for Pij less than or equal to 63

(equation 5) Dij ═ (Pij-Bi) · (63/Mi)

For Pij greater than 63

(equation 6) Dij ═ (Pij-Bi) · (255/Ai)

It will also be appreciated that if the non-linearity of a pixel is similar to a different pixel within the operating range of the system (i.e. the difference in non-linear response is relatively small), correction for the non-linearity is in most cases not required. Non-linear corrections can be ignored if the application does not require a linear response, but only desires to reduce the effects of pixel response non-uniformity across the display frame. All pixel corrections can be ignored if the noise pattern resulting from ignoring all pixel corrections does not interfere with the application. Depending on the application, the correction may be performed at different levels of complexity (linear, bilinear, trilinear, polynomial interpolation, etc.) or not at all.

Variable ROI and variable rotational velocity profile:

in the above examples, different rotation profiles with different rotation speeds are illustrated. In the following example, the rotation profile of the variable speed will be described in the context of an ROI of the image. In the collimator example described above, the central circular region (e.g., 600 of fig. 6 and 700 of fig. 7) serves as the ROI, and thus receives more DPP than the annular sectors 702 and 704 that receive less DPP. This is often the case, and usually the central region of the image is also the ROI, where a less important part of the image is located. Higher DPP results in higher S/N for this region, thus providing better image quality (e.g., better resolvable detail) in that region. Generally, during, for example, a catheterization process, the bed is moved to keep the catheter within the confines of the area 700. However, sometimes the region of most interest in the image moves outside of region 700. Such as the region indicated by reference numeral 1700 in fig. 17A. This may be the result of a variety of reasons, such as (1) the catheter tip moving to region 1700 without the patient moving to bring the catheter tip to region 700, (2) the operator viewing region 1700 for any reason. This new ROI information may be fed back as an input to the system in a number of ways, including automatically following the catheter tip or following the region viewed by the operator using an eye tracking device (e.g., Eyelink 1000 from SR research, inc. of Kanata, ontario, canada) to indicate the desired ROI position to be combined with the user's point of regard, or by using a computer mouse to indicate the desired ROI position.

With the angular span of the aperture sector 702 and the constant rotational speed of the collimator, the DPP of the annular outer region 700 is 1/10 of the DPP of the annular inner region 700, and the S/N of the annular outer region 700 is 1/10 of the S/N of the annular inner region 700^1/2This results in a lower picture quality. To overcome this problem and maintain the refresh rate of 10fps display frames for EC of 1/10 seconds for collimator 500 as in the basic example of the invention, the rotation profile configuration may be modified so that the collimator rotation speed in sector 1702 (fig. 17B) containing region 700 will be reduced to 1/10 for a uniform speed, while the rotation speed in the rest of the EC will be increased to maintain EC of 1/10 seconds.

Reference is now made to fig. 17B and illustrated with an actual numerical example.

Assume that the angular span of sector 1702, which happens to include region 1700, is 54 degrees. The first edge of sector 1702 is 1702A, which is located at an angular position of 63 degrees, while edge 1702B is located at an angular position of 117 degrees. I.e., sector 1700 is centered at a 90 degree angular position.

In this example, when the edge 702A of the sector 702 reaches an angle of 63 degrees (position of 1702A), the rotational speed of the collimator 500 is reduced to 1 rps. This rotational speed is maintained to the position where the edge 702B of the sector 702 reaches the edge 1700B (117 degrees). From this point on, the rotation speed of the collimator 500 increases again. For simplicity, it is assumed that the acceleration and deceleration are very high, and therefore the acceleration and deceleration times are negligible in this example. For the above description, the collimator 500 rotation configuration includes 54+36 at a speed of 1rps being 90 degrees (1/4 for EC rotation). To compensate for this and complete EC at an average 10rps, the rotational speed of collimator 500 must be increased to X rps at 3/4 where EC rotates the remainder, satisfying the following equation:

(formula 7)1rps · 1/4+ X rps · 3/4 ═ 10rps

Thus, it is possible to provide

(formula 8) X rps ═ 10rps-1rps · 1/4)/(3/4)

That is, during the remaining 270 degrees of rotation of the EC, the rotational speed should be 13 rps.

With this rotational configuration, sector 1702 is exposed to the same DPP as region 700, while the S/N of region 1700 is also the same as that of region 700.

It will be appreciated that increasing the collimator rotation speed to 13rps in the sector range outside the sector 1702 will reduce the DPP to below that of the constant rotation speed and is 1/13 of the DPP of the region 700.

It should also be understood that region 700 is presented herein as an example only to explain the design of rotational configurations according to different ROI geometries. The region 1700 may differ in shape and position, and it is possible that multiple ROIs are added to the primary ROI of the circle 700. These changes are handled with configuration changes of the same concept as described above.

It should also be understood that the above described accelerations and decelerations are unreliable parts of EC and must be accounted for. Assume in the next example that acceleration and deceleration each take 45 degrees of rotation and that they are uniform. In this case, acceleration must begin 45 degrees before edge 702A reaches the position of edge 1702A, and deceleration must begin when edge 702B reaches the position of 1702B. All other parameters of the system are the same. If X represents the rotation speed during EC of 180 degrees and Y is the average rotation speed during each 45 degree acceleration/deceleration sector, the following formula must be satisfied to maintain EC of 0.1s (or average rotation speed of 10 rps):

(formula 9)1rps · 1/4+2 · Y rps · 1/8+ X rps · 1/2 ═ 10rps

Assuming that the acceleration and deceleration are constant between 1rps and 10rps, Y ═ 1+10)/2 ═ 5.5, the high rotation during 180 degrees is 16.75 rps.

It should be understood that the method set forth in the above example is also applicable to other acceleration configurations, other collimators, and other modes of operation (e.g., different fps rates). It will also be appreciated that the pixel correction method described above is also fully applicable to different rotational speed configurations.

Different refresh rates for different image areas:

it has been proposed above (using the example of collimator 500 of fig. 5 and an operating mode of 10rps of collimator constant rotation speed and 10fps of display frame refresh rate) that the DPP of the circular region 700 of fig. 7 is 10 times higher than that of the annular region (simply referred to as "ring") formed by sectors 702 and 704. Thus, the S/N of region 700 is also 1/10 better than the S/N of the annular region^1/2. The refresh rate of the entire image 120 (FIG. 2) is the same: 10 fps. The temporal resolution of the entire frame is 0.1 seconds. In the previous example, each display frame was constructed from data from one frame from the camera 116. The area 200 on the display 118 is identical to the area 700 on the sensor. The DPP of region 200 is 10 times higher than the DPP of region 202, and the S/N of region 200 is also 1/10 better than the S/N of annular region 202^1/2. At each EC of the collimator, data is read from the sensor 714 and then processed and displayed on the display 118. The complete image 120 is refreshed every 0.1 seconds.

In the following example of the invention, it is desirable to improve the S/N of ring 202.

In the first example, the ring 202 is refreshed only once every 1 second, although the region 200 is refreshed once every 0.1 seconds and with data read from the sensor 714. During this 1 second period, the data received from the sensor 714 for the pixels of the ring 202 is used to generate a ring image of the sum of the previous 10 frames. In a simplified form, all 10 frames with an index j of 1 to 10 are stored. Then for each pixel i in the range of the ring 202, the sum of the values is calculated: pni ═ Σ pij. Then corrected and displayed Pni, where n is the index number for each 10 frame group. Thus for j 1 to 10, the pixel of the sum frame is P1 i. For j equal to 11 to 20, the pixel of the sum frame is P2 i. For j 21 to 30, the pixel of the sum frame is P3i, and so on. Thus, using this example, a display of image 120 is obtained in which the S/N of loop 202 is similar to that of region 200, although loop 202 receives 1/10 for the DPP of region 200 at each unit time. It is to be compromised that the ring 202 is refreshed every 1 second relative to the region 200 every 0.1 second, and that the temporal resolution of the ring 202 is 1s relative to the 0.1 second temporal resolution of the region 200.

In a second example, the ring 202 is refreshed in a different manner after the first 10 frames with indices j 1 to 10 are obtained and stored and displayed as the sum of the pixels of the ring 202. Instead of maintaining the display of the ring 202 for 1 second until j is obtained 11 to 20, the displayed image is refreshed after 0.1 second as follows:

frame j is obtained 11 and stored in place of frame 1. Thus previously stored frames 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, later stored frames: 11. 2, 3, 4, 5, 6, 7, 8, 9, 10. The set of frames is processed in the same manner as the previous set and the ring 202 is refreshed. After another 0.1 seconds, a frame with index 12 is obtained and stored to replace the frame with index 2: 11. 12, 3, 4, 5, 6, 7, 8, 9, 10. The group is processed in the same manner and the display of the ring 202 is refreshed. This process is repeated so that the annular region is refreshed every 0.1 seconds, as is region 200. The temporal resolution of the ring 202 is still 1 second, relative to the 0.1 second temporal resolution of the region 200. The S/N of ring 202 is similar to the S/N of region 200.

In a third example, an intermediate approach is proposed. After the first example, instead of summing 10 frame pixels and refreshing the ring 202 every 1 second, summing every 5 frames and refreshing the ring 202 may be every 0.5 seconds. The S/N of ring 202 will now be 1/2 of the S/N of region 200^1/2But still greater than 1/10 of the basic example of collimator 500^1/2Good and the time resolution is only 0.5 seconds, relative to 1 second for the first example of the method.

It should be appreciated that an intermediate approach may also be used in the second example, and that one of the groups of 5 frames may be replaced instead of replacing each timeOne of 10 frames, namely: 1. 2, 3, 4, 5, then 6,3, 4, 5, and so on. Here again a refresh of ring 202 every 0.1 seconds is obtained, but a temporal resolution of 0.5 seconds, and the S/N of ring 202 will now be 1/2 of the S/N of region 200^1/2But still better than 1/10 in the basic example of collimator 500^1/2Good results are obtained.

It will be appreciated that this method may also be used for collimators other than rotary collimators, such as the one in figure 18. FIG. 18A provides a top view of the collimator and FIG. 18B is a cross-sectional view of c-c of FIG. 18A. Collimator 1800 provides similar functionality as the X-ray reduction of the collimator of the present invention. Which has an aperture 1802 to allow all radiation in that area to pass, a ring 1806 to reduce the amount of radiation passing through that area depending on the material (usually aluminum) and material thickness, and a ring 1804 having a thickness that varies as a function of distance from the center, starting at zero thickness on the aperture 1802 side and ending at the ring 1806 thickness on the ring 1806 side. Fig. 18C provides a schematic DPP chart as a function of distance r from the center.

The above assumes that radiation outside of ring 1806 is completely blocked. For purposes of illustrating this example, radiation scattered from collimator 1800 is ignored. For the present example, assume that the DPP through ring 1806 is 1/10 of the DPP through aperture 1802. The frame rate is 10fps and the display frame refresh rate is 10/sec. As described in the example above, the S/N of the image portion associated with ring 1806 is 1/10 of the S/N associated with aperture 1802^1/2. To display an image in which the S/N of the area associated with ring 1806 is similar to the S/N of the area associated with aperture 1802, any of the above-described methods may be used.

FIG. 18D provides a presentation of display 118 with display frames associated with collimator 1800. Circle 1822 is the area associated with radiation arriving through aperture 1802 of collimator 1800. Ring 1824 is a region associated with radiation arriving through ring 1804 of collimator 1800. Ring 1826 is the region associated with radiation arriving through ring 1806 of collimator 1800. It should be appreciated that while the thickness of the example ring 1804 in FIG. 18B varies linearly, the example radial variation of 1814 in FIG. 18C is a non-linear thickness variation. That is, many different functions may be used to produce the slope of the thickness 1804 to suit the desired step change in the amount of radiation between the rings 1800 and 1806 of FIG. 18B.

In the first example, the ring 1826 is refreshed only once every 1 second, although the region 1822 is refreshed every 0.1 seconds and the sensor 714 is used to read the data. During this 1 second, the data received from sensor 714 for the pixels of ring 1826 is used to generate a ring image that is the sum of the previous 10 frames. In a simplified form, all 10 frames with an index j of 1 to 10 are stored. Then for each pixel i in the range of ring 1826, the sum of the values is calculated: pni ═ Σ pij. The correction and display Pni is then made, where n is the index number of each group of 10 frames. Thus for j 1 to 10, the pixel of the sum frame is P1 i. For j equal to 11 to 20, the pixel of the sum frame is P2 i. For j 21 to 30, the pixel of the sum frame is P3i, and so on. Thus, using this example, a display of image 120 is obtained in which the S/N of ring 1826 is similar to the S/N of region 1822, although ring 1826 receives DPP 1/10 of region 1822 at each unit time. To compromise, the ring 1826 refreshes every 1 second relative to the region 1822 every 0.1 second, and the ring 1826 has a temporal resolution of 1s relative to the region 1822's temporal resolution of 0.1 seconds.

For ring 1824, an example of 1/10 where the DPP linearly decreases from the DPP of 1822 to this DPP (the DPP of ring 1826) across the width of ring 1820 may be used.

In this example, ring 1824 may be divided into 8 rings of the same radius, such that the average DPP in the smallest ring #1 is 9/10 of 1822, the DPP in the next ring #2 is 8/10 of 1822, ring #3 is 7/10, and so on, until the last ring #8 is 2/10 of the DPP of 1822.

Whenever a value is mentioned with reference to the above-mentioned sections (ring #1 to ring #8), the value is the average value of that section, taking into account the thickness variation of the collimator through that section. When the objective is to provide the same S/N in the complete display image 120 and to maintain the time resolution up to 1 second, the pairing of ring #5 (1/2 for DPP in region 1822) and ring #8 (1/5 for DPP in region 1822) can be done in a simple manner, since the ratio of DPP in region 1822 to DPP in ring #5 is an integer. The same is true for ring # 2.

In the case of ring #5, adding 2 temporally consecutive frames as described in any of the above methods (with sufficient pixel correction as described above) will provide an S/N similar to region 1822. The time resolution in this example is 0.2 seconds.

In the case of ring #8, adding 5 temporally consecutive frames as described in any of the above methods (with sufficient pixel correction as described above) will provide an S/N similar to region 1822. The time resolution in this example is 0.5 seconds.

For the other rings (#1, #3, #4, #6, #7, and #8), the ratio of the DPP in region 1822 to the DPP in any of these rings is not an integer. So increasing the pixels for an integer number of frames (considering the desired limit of up to 10 for a temporal resolution of no more than 1 second) will exceed the desired S/N or be less than the desired S/N.

To achieve the desired S/N with the requirements of this example, the following method may be used:

1. for each ring # m, the minimum number of pixels of the temporally successive frame is added, which provides an S/N equal to or higher than that of the region 1822.

2. Performing pixel correction (shifting, normalization, etc. as described above)

3. Noise is added to each pixel in ring # m to compensate for the fact that the S/N is higher than in region 1822.

The above steps will be described in more detail below with reference to ring # 1.

The DPP in loop #1 is 9/10 of the DPP of region 1822. The S/N in ring #1 is that of region 1822 (9/10)^1/2. Therefore, according to the above step 1, it is necessary to add pixels of two temporally successive frames in the region of the ring #1 so that the ring #1The S/N of the middle pixel is equal to or greater than the S/N of region 1822.

By adding two temporally consecutive frames of pixels in the region of ring #1, the effective DPP of the final frame obtained in ring #1 is 18/10 of the DPP of region 1822. The S/N in ring #1 is that of region 1822 (18/10)^1/2。

To compensate for the excessively high S/N (and thus possible visual artifacts in the image 120), Gaussian noise is added to each pixel to satisfy the formula:

(equation 10) (N)₁₈₂₂)²＝(N_#1)²+(N_add)²

Where N1822 is the noise associated with a particular pixel in region 1822 transmitted for a particular object, N1822_#1Is the noise associated with the pixels (pixel sums) of the 2 temporally successive frames of ring #1 that have the same object transport and after the pixel sums have passed through the pixel correction process (which in its simplest form of correction involves dividing the pixel sums by 1.8 to bring the effective DPP from 18/10 down to 10/10-the same as in region 1822), N_addIs noise added to the pixel sum to bring its S/N to the same level as the equivalent pixels in region 1822.

In the above example, since the number of X-ray photons is 1.8 times the number of identical pixels (same object transmission) of region 1822 in the sum of pixels of ring #1, the noise of the sum of pixels is (1.8) of identical pixels of region 1822^1/2And S/N is also equal pixel of region 1822 (1.8)^1/2。

To calculate N_addIn the following form, we use equation 10:

(formula 11) N_add＝((N₁₈₂₂)²-(N_#1)²)^1/2

Divide by 1.8 with pixel correction.

Using the numbers:

N_add＝(1²-((1.8)^1/2/1.8)²)^1/2

N_add＝0.667

thus, by adding this poisson noise to the pixel sum, noise similar to that of the equivalent pixels in region 1822 is provided to that pixel.

It should be understood that all examples are calculated on a relative basis, and thus the pixel of region 1822 is 1.

It should be understood that the noise value in equation 10 depends on the pixel value, typically the square root of the average level of the pixels.

The same correction method can be used for all portions of ring 1824 using appropriate adjustments.

It will be appreciated that adding pixels for successive frames can be done by refreshing the display frame or adding a new frame each time or using a FIFO method as described above.

It should be understood that the division of the ring 1824 into 8 sections (ring #1 through ring #8) is provided as an example only. The greater the number of sections, the more uniform the S/N of the ring 1824 will be. However, the visibility of the S/N adjustment unevenness is masked by the S/N of the image, and therefore, when a certain number of portions are exceeded, the visual contribution of more portions is low and may not be distinguishable by the operator. The number of loop portions can thus be limited in a particular procedure based on S/N statistical analysis of the image.

The same method of handling areas of inhomogeneous DPP, such as ring 1824 of collimator example 1800, may also be used for collimators of the present invention that also produce areas of inconsistent DPP, such as all collimators of fig. 15C, 15D, and 16. These methods can be used for any collimator that produces different exposure areas, regardless of the method used by the collimator, e.g. whether the different exposure areas are produced by collimator shape, collimator movement or a combination of shape and operation. In all cases of collimator movement, the same period of movement pattern as described above simplifies the image enhancement, but it is not necessary to allow the image enhancement described above.

In another example of the present invention, when the ROI of region 1700 as shown in fig. 17A is shifted, instead of adjusting the rotational configuration of collimator 500 as described with reference to fig. 17B, the entire collimator may be linearly shifted in a direction parallel to the plane of collimator 500, such that the X-ray radiation passing through circular aperture 504 of fig. 5 is now concentrated around region 1700, as shown in fig. 19A on camera sensor 710. It is assumed that the only rays that may reach the collimator input surface 112 are the rays that pass through the aperture of the collimator 500 (circular aperture 505 and fan-shaped aperture 506). Thus, region 1902 in the sensor is obscured in FIG. 19A (no rays reach the corresponding region of the image intensifier input 112), and only the region including 700, 702 and 704 bounded by boundary 712 is exposed. The exposed areas then overlap between the two circles and the centers of one are offset relative to the other, as indicated in fig. 19A by the numerical designation 1900.

This desired functionality of the present invention is provided herein in area 1900 by circular aperture 504 which enables high DPP in area 700 and fan-shaped aperture 506 which enables DPP only for aperture 504 1/10 associated with the rest of the image area.

Fig. 19B shows the display version of fig. 2 according to the example of fig. 19A.

Collimator 500 may be moved in the X-Y plane using any common X-Y mechanical system (see coordinate system 126 of fig. 1). For example, the annular stationary portion 1016 of FIG. 10C is connected to an X-Y system rather than to an X-ray tube structure, while the X-Y system is connected to an X-Y tube structure, thus enabling the collimator of FIG. 10C to move in the X-Y plane in this example, as described in the example of FIG. 19A.

It will be appreciated that the above-described methods, such as pixel correction, S/N adjustment, adding pixels for different frames, are fully applicable to the example of adjusting the shift of the collimator of fig. 19A. The X-Y shift method can be used for any collimator of the present invention.

It will be appreciated that shifting along a straight line (e.g., along the X-axis) rather than X-Y may be used in the same manner, with the limitation that regions of the ROI may be processed in this manner in the region of the image 120.

The X-Y mechanical system may have many different designs including, for example, motorized XY table ZXW050HA02 from shanghai orthoxin corporation, shanghai, china. Custom design of X-Y mechanical systems is common in the art and is often tailored to meet the needs of a particular application. One such provider of custom designed X-Y mechanical systems is the LinTech by Monrovia, California, USA.

It should be appreciated that the diameter of collimator 500 may be increased such that the length of sector 702 is increased to r3, as shown in fig. 20B.

Fig. 20A is the collimator of fig. 5 provided herein as fig. 20A for simple comparison with the collimator of fig. 20B. The angles 508 are the same (36 degrees in this example) and the diameter of the circular aperture 504 is the same (r 1). R3 is large enough to incorporate the full field of view of the image intensifier input 112, for example, when the collimator is moved laterally as described with reference to FIG. 19. With this design, the full image area 120 of FIG. 19B remains active without any shadow (obscured) areas, such as 1902 in the example of FIG. 19. This collimator enlargement may be implemented in any of the collimators of the present invention.

For the example of FIG. 19, where the desired maximum displacement is the point at which the edge circular aperture 700 contacts any point on the edge of the image 712 (e.g., the point 1904 in FIG. 19A), the desired radius r3 of the sector aperture may be calculated as follows with reference to FIG. 20B:

(equation 12) r3 ═ a-r 1

Where A is the diameter of the image intensifier input 112B (see FIG. 3) projected proportionally onto the collimator plane:

(equation 13) A. B. (D1/D2)

During movement of the collimator in the X-Y plane, pixels that have been completely DPP exposed (via area 504) may change state to DPP exposed at 1/10 because area 504 has moved and these pixels are no longer included in that area. It should be appreciated that in 1 second, the state of the pixel changes from being included in region 504 and being completely DPP to being outside region 504 and 1/10DPP, 10 frames of 1/10DPP have been obtained, and the processing of this pixel for display has been completed in any of the methods described above using the last 10 frames (or 5 frames after 0.5 seconds in another example) to provide the same S/N as in region 504. During this 1 second transition, another process is required to keep the S/N of this pixel the same as it was included in region 504. In this example, the following process is implemented with a refresh rate of 0.1 seconds and a temporal resolution that varies from 0.1 seconds to 1 second, where N is the index of the last full DPP frame for that pixel:

1. at time 0, 100% of the pixels of the last full DPP data for frame N are displayed. The temporal resolution is 0.1 seconds.

2. At time 0.1 seconds, pixels are displayed that are a weighted sum of 90% of the last full DPP data for frame N and 100% of the new DPP data for frame N + 1.

3. At time 0.2 seconds, pixels are displayed that are a weighted sum of 80% of the last full DPP data for frame N, 100% of the DPP data for frame N +1, and 100% of the DPP data for frame N +2.

4.….

5.….

6.….

7.….

8.….

9.….

10. At time 0.9 seconds, pixels are displayed that are a weighted sum of 10% of the last full DPP data for frame N and 100% of the new DPP data for each frame N +1, N +2, …, N + 9.

11. At time 1.0 seconds, pixels are displayed that are a weighted sum of 0% of the last full DPP data for frame N and 100% of the new DPP data for each frame N +1, N +2, …, N +9, N + 10. The temporal resolution now changes to 1 second.

12. The above method is continued to improve the image for the 1/10DPP region. The temporal resolution is 1 second.

It should be appreciated that in the case of the method where 1/10DPP pixels are refreshed at a rate of only 1fps, the last full DPP data will appear for 1 second after the pixel is changed to 1/10DPP exposure, and then the average of the last 10 1/10DPP frames will be used to refresh the pixels.

In the case of a pixel changing state in the opposite direction, i.e. from 1/10DPP area to full DPP area, the transition is instantaneous and the displayed image is refreshed with the first 0.1 second hand of full DPP the first 0.1 second after the state change.

It will be appreciated that the method described above may also be used for pulsed X-rays of relatively high frequency, as described with reference to figure 1. The term "relatively high frequency" is with respect to the collimator design and mode of operation. In the collimator 500 example of fig. 5, with a sector angular span of 36 degrees and a rotation of 10rps, the pulse frequency should be at least at the frequency 100/s, so that there is at least one X-ray pulse in each 36 degree region of the frame. To simplify the pixel correction approach, it is also desirable that the X-ray pulse frequency should be a positive integer multiple of the minimum frequency. In this example: 200/s, 300/s, 400/s, and so on. In this example, 1000/s (10 times the minimum frequency) may be considered a relatively high frequency.

It should be understood that none of the collimators is completely opaque to X-rays, whereas the collimators are configured to block most of the X-rays in the opaque portion. With a 0.25mm (lead-like) HVL (half value layer), a 3mm thick collimator would allow 0.5 of the incident X-ray radiation^(3/0.25)1/4096 pass (no scatter). The term "substantially opaque" will be used to describe these actual collimators. Most collimators described herein consist of substantially opaque regions such as 518 and apertures or holes (e.g., 504 and 506 of FIG. 5) of FIG. 5Is constructed. The collimator example of fig. 18, for example, is different because they include a semi-transmissive region (e.g., 1804 of fig. 18A) in addition to the substantially non-transmissive region 1806 and the aperture 1802.

The collimator according to the invention can be fitted to the X-ray system, for example, independently or together with another collimator, so that it is designed to limit the X-rays from reaching a part of the input region 112 of the image intensifier.

The collimators of the present invention and the other collimators may be placed in any order along the X-ray path. The exposed portion of region 112 will be the remainder of the overlapping area of all collimators in the X-ray blocking path. In this sequentially arranged design, the distance of each collimator from the X-ray source and the distance to the area 112 need to be considered in conjunction with the collimator geometry, as described above, to achieve the desired functionality.

Those skilled in the art will appreciate that the methods and techniques described above are not intended to limit the configurations and methods mentioned herein as examples. These are provided as examples only, and other configurations and methods may be used to optimize the end result depending on the particular design and set of techniques implemented in the production of the design.

The above embodiments are described herein by way of example only and do not limit the scope of the present invention.

The scope of the invention is to be determined entirely by the claims provided herein.

Claims (5)

1. An X-ray system comprising an X-ray source, a single substantially circular collimator, a detector and a display, means for moving the collimator in a plane substantially parallel to the plane of the collimator;

and said collimator comprising a central aperture allowing all radiation to pass, an outer ring for reducing the amount of radiation passing according to the material and the thickness of the material, and an inner ring between said central aperture and said outer ring, the thickness of said inner ring varying as a function of the distance from said central aperture, starting at zero on the side of said central aperture and ending at the thickness of said outer ring on the side of said outer ring;

wherein the detector is configured to capture X-ray radiation, the X-ray system further configured to read frames comprising pixels from the detector after each frame is captured by the detector;

the X-ray system is configured to calculate a gain per frame and an offset correction per frame, and calculate a normalization factor per frame based on a different per-pixel radiation dose for each of the central aperture, the outer annulus, and the inner annulus of the collimator.

2. The X-ray system of claim 1, configured to calculate the normalization factor for the inner annular ring by dividing the inner annular ring into a plurality of rings and using a single dose per pixel estimate to each of the plurality of rings based on distance from the central aperture.

3. The X-ray system of claim 1, further comprising an eye tracker configured to track an operator's gaze, thereby determining a location of a region of interest and moving the collimator accordingly.

4. A method for enhancing a displayed exposure image in an X-ray system comprising an X-ray source, a single substantially circular collimator, a detector and a display, means for moving said collimator in a plane substantially parallel to the plane of said collimator; wherein the collimator comprises a central aperture allowing passage of all radiation, an outer ring for reducing the amount of radiation passing according to the material and the thickness of the material, and an inner ring between the central aperture and the outer ring, the inner ring having a thickness varying as a function of the distance from the central aperture, starting at zero on the side of the central aperture and ending at the thickness of the outer ring on the side of the outer ring, the method comprising:

acquiring a frame comprising pixels from said detector;

calculating a gain for each frame and an offset correction for each frame; and

calculating a normalization factor for each frame based on a different dose per pixel for each of said central aperture, said outer annulus and said inner annulus of said collimator;

wherein said calculating a normalization factor for said inner circle comprises dividing said inner circle into a plurality of rings and assigning a theoretical value of dose per pixel to each of said plurality of rings based on distance from said central aperture.

5. The method of claim 4, further comprising tracking the operator's gaze, thereby determining the location of the region of interest and moving the collimator accordingly.